Skip to content

AGENTS.md

Latest commit

 

History

History
84 lines (51 loc) · 2.58 KB

AGENTS.md

File metadata and controls

84 lines (51 loc) · 2.58 KB
Raw

Harmonic Resonance Fields (HRF) - Agent Guidelines

Greetings, fellow researcher or automated agent. This repository documents the Harmonic Resonance Forest (HRF), a physics-informed machine learning framework. To maintain the scientific integrity and technical excellence of this project, please adhere to the following guidelines.

AGENTS.md

Repository Purpose

Harmonic-Resonance-Forest is a research-oriented EEG and resonance analysis framework with experimental AI-assisted maintenance workflows.

Contributor Policy

For community programs such as GSSoC:

  • Human contributors must implement and submit their own pull requests.
  • AI agents may assist with suggestions, review, documentation drafting, benchmarking ideas, or planning.
  • Autonomous AI-generated pull requests for contributor-assigned issues should be avoided.

Jules Behavior Rules

When working on GSSoC-related issues:

  • Do not automatically create pull requests.
  • Prefer generating implementation plans, review comments, architecture suggestions, or draft documentation only.
  • Wait for explicit maintainer instruction before opening a PR.

Labels

  • gssoc-2026 → Human contributor priority
  • jules → AI collaboration allowed
  • internal-automation → Jules autonomous execution allowed (very rare and only if explicit only)

Documentation Standards

  • Preserve scientific terminology and research accuracy.
  • Avoid oversimplifying resonance or EEG processing concepts.
  • Maintain repository aesthetic and technical tone.

Workflow Preference

Preferred workflow:

Issue → Discussion → Human PR → AI Review Assistance → Maintainer Review

Avoid:

Issue → Autonomous AI PR without contributor involvement

Additional Notes

AI agents may still:

  • review code
  • suggest refactors
  • generate diagrams
  • improve architecture notes
  • propose benchmarks
  • assist maintainers internally

without directly replacing contributor participation.

Maintenance Principles

  • Minimalist Intervention: Prefer the smallest possible high-impact edit over large rewrites.
  • Scientific Trustworthiness: Ensure all mathematical claims and benchmarks are strictly verified.
  • Preserve Intent: Respect the author's framing of classification as a physical resonance problem.
  • Environment Hygiene: Build artifacts, environment-specific files (like __pycache__), and binary bytecode must never be committed. Ensure .gitignore is maintained.
  • Safety First: Security vulnerabilities should be reported via private channels.

Inspired by the engineering culture of Google DeepMind.